[Adium-devl] msn-pecan status update?

[Felipe Contreras]

Hi,

On Tue, Jul 29, 2008 at 2:28 AM, Evan Schoenberg <evan.s at dreskin.net> wrote:
>
> On Jul 28, 2008, at 3:09 PM, Felipe Contreras wrote:
>
> First of all I need to know which versions of msn-pecan correspond to
> your version of Adium; I see reports of 1.3b9 but I don't know the
> msn-pecan version, similarly with svn builds. If I could find that out
> from the svn code that would certainly help.
>
> 1.3b9 is [24466] (per the About Adium window, beta.adiumx.com, or svn log
> adium/Plists)
> svn log adium/Frameworks/libpurple.framework/ shows that libpurple.framework
> last changed msn-pecan at [24436]: im.pidgin.pidgin 2.5.0mtn @ 2dfa29ca with
> msn-pecan 0.0.14-63-g23fd241.
> libpurple.framework's commit messages should always contain the
> im.pidgin.pidgin and msn-pecan revisions unless there was no change (e.g.
> the series of commits which were me iterating on fixing an ICQ issue and
> providing builds to testers).

Thanks, that definitely helps.

> I know it sounds pedantic but many issues are impossible to fix
> without the right information:
> * version of msn-pecan
> * debug log
>
> Definitely not pedantic. Very reasonable.

Hopefully the users realize that :)

> In the case of crashes:
> * backtrace of the crash *with* debugging symbols [1]
>
> Running in gdb gets you a backtrace; it doesn't add debugging symbols
> unless the person has compiled his/her own libpurple.framework and kept the
> object files around.  So far as I can tell this is the only way debugging
> symbols can stay around; I see no way to include them in the
> libpurple.framework binary itself.
> If anyone can figure out a way to keep the debug symbols in the binary, it'd
> definitely be worth the space tradeoff to do that for our beta builds.

Usually the symbols are kept, so something in the process must be
striping the symbols.

>  Alternately, if we can get a DWARF file to actually be created-as-requested
> and then distribute that to testers, that'd work, too... assuming that one
> can tell gdb where to go looking for the file.

This is described in:
http://sources.redhat.com/gdb/onlinedocs/gdb_16.html#SEC156

> I'd also like to point out #10411 which you didn't mention in your ticket
> list; it appears to be a problem new to msn-pecan, at least based on the
> reports, and it's a significant issue.

Right, probably I will be able to reproduce this one, so yeah, I'll fix it soon.

-- 
Felipe Contreras